JP2002169810A - Computer-readable recording medium with recorded image retrieval program, and method and device for image retrieval - Google Patents

Abstract

PROBLEM TO BE SOLVED: To retrieve a desired image without any omission while a user is not forced to perform complicated operation. SOLUTION: When the user specifies a retrieval object period and a retrieval key image (S601, S603), images in the past before time information added to the retrieval key image are repeatedly retrieved (S607 to S613) while traced back at intervals of a specific period T1. Each time the retrieval period shifts, the retrieval key image is also updated to the oldest image among last retrieval results. When the retrieval period exceeds the lower limit Tmin of the retrieval object period, images in the future after the time information of the retrieval key image are repeatedly retrieved (S617 to S623) while advanced at intervals of the specific period Td. Here, each time the retrieval period shifts, the retrieval key image is updated to the newest image among last retrieval results. When the retrieval period exceeds the upper limit Tmax of the retrieval object period, the retrieval is completed and the retrieval results in the retrieval object period are outputted (S625).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer-readable recording medium storing an image search program, an image search method, and an image search apparatus. The present invention relates to a computer-readable recording medium storing an image search program capable of automatically setting a search key image and performing a search again with the new search key image, an image search method, and an image search device.

[0002]

2. Description of the Related Art An image photographed by a digital camera or an image read by a scanner is taken into a computer for output and editing. Then, the captured image may be provided with necessary information such as a shooting date and time and may be stored in a database. The following method is known as a method for searching for a necessary image from images in a database.

First, a user designates an image serving as a search key (hereinafter referred to as a "search key image").
When a search key image is specified, an image similar to the search key image is extracted from the images stored in the database by taking a difference between the search key image and the image stored in the database. Is done. Then, the extracted image is returned to the user as a search result.

[0004] By using such an image search method,
A desired image can be easily selected from a database in which a huge number of images are stored. For example,
When the user wants to search for an image in which he or she is reflected, the user specifies his / her own face image as a search key image. Then, all images similar to the search key image are extracted from the database.

[0005]

However, according to the above-described image search method, since an image is searched only by a search key image specified by a user, it is difficult to extract a desired image without omission. There was a problem.

That is, for example, in the case of a person, changes in age, body shape, hairstyle, and the like occur over time. Therefore, when an image taken at a certain time is designated as a search key image, the same person is likely to be accurately extracted from images whose shooting time is close to that image, but the shooting time is different In some cases, the same person may be determined to be another person due to changes in age, body shape, hairstyle, and the like.

[0007] For this reason, the user, in anticipation of such a search result, specifies a plurality of representative search key images in advance, or considers the tendency of the image extracted by the search key image specified first. Then, it is necessary to specify a new search key image and perform the search again. That is,
To extract a desired image from the database without omission, the user had to perform some complicated operation.

The present invention has been conceived in view of the above circumstances, and has as its object to provide an image retrieval program that enables a user to extract a desired image without omission without complicated operations. It is an object of the present invention to provide a computer-readable recording medium, an image search method, and an image search device.

[0009]

According to an aspect of the present invention, there is provided a computer-readable recording medium comprising: a retrieval key image acquiring step for acquiring a retrieval key image designated by a user; A search step of searching for a corresponding image from a database based on the obtained search key image, a setting step of setting a new search key image using a search result of the search step, An image search program for causing a computer to execute a re-search step of re-searching the corresponding image from the database based on the search key image.

According to the present invention, when the user specifies a search key image, the corresponding image is searched from the database based on the search key image. Then, a new search key image is automatically set using the searched result, and the corresponding image is searched again based on the new search key image.

For this reason, the user who desires the search only has to specify the search key image once for the search at first, and then does not need to reset the search key image. Also, since a new search key image is set using the result based on the search key image specified first,
Even when an image to be searched for changes with the passage of time or the like, a desired image considering the change is appropriately extracted.

Therefore, it is possible to provide a computer-readable recording medium on which an image search program capable of extracting a desired image without any omission without a user's complicated operation is recorded.

[0013] Preferably, the search step performs a search on a first group in the database, and the re-search step includes a search on a second group different from the first group in the database. Perform

According to the present invention, first, the first group is searched based on the search key image specified by the user, and then the second group different from the first group is searched based on the new search key image. The search is performed for the group of. Since an appropriate group is searched according to the search key image, images desired by the user are extracted without omission.

Preferably, each of the images stored in the database is provided with time information.
Is a group for images having time information in the first period, and the second group is a group for images having time information in the second period.

According to the present invention, an image stored in the database is provided with time information such as a photographing date and a registration date of the image. Then, a search is performed for an image having time information within the first period based on a search key image specified by the user, and a search is performed for an image having time information within the second period based on a new search key image. Done. Since images having different periods such as shooting dates are searched based on the respective search key images, it is possible to perform an appropriate search in consideration of a change in a search target due to a change in time.

Preferably, in the setting step, an image having time information close to the second period is set as a new search key image from the images searched in the search step.

According to the present invention, the second image is selected from among the images searched using the search key image specified by the user.
An image having time information close to the period is set as a new search key image. Since an image whose shooting date is close to the second period is set as a new search key image, it is possible to perform a more appropriate search for an image shot or the like during the second period.

Preferably, the computer-readable recording medium further includes a search target period obtaining step of obtaining a search target period specified by a user, and an image obtained by the search step and the re-search step. And a determining step of determining an image having time information within the search target period as a search result.

According to the present invention, the search target period is designated by the user. Then, an image having time information within the search target period is determined as a search result from the images searched by the search step and the re-search step.
Since the search target period is clearly specified, the image desired by the user is more appropriately searched.

Preferably, the computer-readable recording medium further records an image search program for causing a computer to execute a reset step of setting a new search key image using a search result obtained by the search step. , The resetting step and the re-searching step are repeated a predetermined number of times.

According to the present invention, a new search key image is reset using the result of the re-search based on the new search key image. Then, a re-search is performed using the reset new search key image. Such resetting of a new search key image and re-searching for an image using the set search key image are repeated a predetermined number of times. Therefore,
For example, even when the search range is wide, or when the search key image specified first by the user and the image to be searched for change over time or the like, the user can select the desired image. Extraction can be performed without omission.

According to another aspect of the present invention, an image search method includes: a search key image obtaining step of obtaining a search key image specified by a user; A search step of searching for an image to be searched; a setting step of setting a new search key image using a search result of the search step;
A re-search step of re-searching the database for a corresponding image based on the set new search key image.

According to the present invention, when a user designates a search key image, a corresponding image is searched from a database based on the search key image. Then, a new search key image is automatically set using the searched result, and the corresponding image is searched again based on the new search key image.

For this reason, the user who desires the search only needs to specify the search key image once for the search at first, and then does not need to reset the search key image. Also, since a new search key image is set using the result based on the search key image specified first,
Even when an image to be searched for changes with the passage of time or the like, a desired image considering the change is appropriately extracted.

Therefore, it is possible to provide an image search method capable of extracting a desired image without omission without a user performing a complicated operation.

According to still another aspect of the present invention, an image search device includes a search key image obtaining means for obtaining a search key image specified by a user, and a search key image from a database based on the obtained search key image. Using search means for searching for the corresponding image and search results by the search means,
Setting means for setting a new search key image; and re-search means for re-searching a corresponding image from a database based on the set new search key image.

According to the present invention, it is possible to provide an image retrieval apparatus capable of extracting a desired image without omission without a user performing a complicated operation.

[0029]

Embodiments of the present invention will be described below in detail with reference to the drawings.

An image search method according to an embodiment of the present invention is realized by executing an image search program on a computer such as a personal computer or a workstation.

FIG. 1 is a diagram showing the appearance of a computer which is an example of the image search device 100 according to the present embodiment. With reference to FIG.
1, a magnetic tape device 13, and a CD-ROM (Compact
Disc-Read Only Memory (LCD) device 17, display device 12 such as a CRT, keyboard 15, mouse 16, and modem 19. The magnetic tape device 13 has a magnetic tape 14 mounted thereon, and the CD-ROM device 17 has a CD-R
The OM 18 is mounted.

FIG. 2 shows the configuration of this computer in the form of a functional block diagram. Referring to FIG. 1, as is well known, a computer main body 11 includes a CPU (Central Process
ng Unit) 20 and ROM (Read Only Memory) 21
, A RAM (Random Access Memory) 22, and a hard disk drive 23. These are mutually connected by a bus.

The image search program may be installed in advance on a hard disk in a computer, or may be recorded on a removable recording medium such as a CD-ROM or a magnetic tape. In any case, the image search program is recorded on a computer-readable recording medium.

When the program is recorded on a removable recording medium, the recorded program is stored in the magnetic tape device 1.
3. The data is read from the recording medium by the CD-ROM device 17 or the like and temporarily stored in the hard disk device 23. Thereafter, the program is loaded from the hard disk device 23 to the RAM 22 in advance, and the CPU 20 controls the execution of the program.

As a computer-readable recording medium, a tape system such as a magnetic tape or a cassette tape, a magnetic disk (a flexible disk, a hard disk device, etc.) and an optical disk (CD-ROM / MO / MD)
/ DVD, etc., a card system such as an IC card (including a memory card) or an optical card, or a mask ROM, EPROM, EEPROM, flash RO
A medium that fixedly holds the program, such as a semiconductor memory such as M, may be used.

Further, the medium may carry the program in a fluid manner so that the program can be downloaded from the network via the communication modem 19. In addition,
When the program is downloaded from the network as described above, the download program is stored in the computer main body 11 in advance, or is installed in the main body 11 in advance from another recording medium or the like.

The content stored in the recording medium is not limited to a program but may be data.

Next, referring to FIG.
The case where an image is input / output to / from the computer will be described.

Referring to FIG. 3, control unit 300 corresponds to computer main body 11 of image search apparatus 100 shown in FIGS. The database 301 may be a memory in the computer main body 11 or an external memory. In the database 301, an image photographed by the digital camera 31, an image read by the scanner 32, an image acquired from the server 33 via a network, and the like are registered. In addition, other PC
Alternatively, an image acquired from a recording medium such as a CD-ROM may be registered.

The search key image specified by the user is
The image may be designated from among images registered in the database 301 in advance, or the digital camera 31
The image may be an image photographed by the scanner, an image read by the scanner 32, or an image acquired from the server 33 or the like.

Upon receiving the search key image input by the user, the control unit 300 searches the image stored in the database 301 for a corresponding image. Further, the search key image is automatically updated according to a predetermined condition, and the image in the database 301 is searched again. Such re-search is repeated as necessary, and the image obtained as a result is output as a search result.

As an output destination, a printer 37, a display 38 (12), and a monitor 309 such as a TV can be considered. Further, it may be a PC 34 or a server 35 connected to a network. Further, the images may be distributed to the electronic album 36 or the like.

FIG. 4 shows the control unit 30 of the image search apparatus 100.
9 is a flowchart showing the flow of processing in the process No. 0.
As shown in the figure, first, in step S401, initial settings such as display of an initial screen are performed. Next, in step S403, the menu is selected by the user. For example, a menu for any desired processing is selected based on a selection screen on which image data registration processing, image data search processing, and other menus are displayed.

If the image data registration process is selected, a process for registering the image data in the database 301 is performed in step S405. If the image data search process is selected, step S407
In, a predetermined image search process based on the search key image is performed. If any other menu is selected,
In step S409, other processing such as image processing is performed. And when the selected process is completed,
The process returns to step S401 again.

Next, image data registration (step S40)
Each processing of 5) and image data search (step S407) will be described in detail.

FIG. 5 is a flowchart showing details of the image data registration process (step S405) of FIG.
As shown in the figure, first, in step S501, input of image data is performed. That is, image data (image file) specified by the user or image data transferred from an input device such as the scanner 32 is input to the image search device 100.

Next, in step S503, shooting date information is extracted from the input image data, and given to the image data as one of the key information of the image. In addition,
A registration date may be given as key information instead of a shooting date.

Subsequently, in step S505, other key information is added to the corresponding image data. The other key information is, for example, a keyword input by the user. When the input of the image data and the input of the necessary key information are completed, the input image and the key information are registered in the database 301 in association with each other in step S507.

FIG. 6 is a flowchart showing details of the image data search process (step S407) of FIG.
First, in step S601, the user specifies a period to be searched (hereinafter, referred to as a "search target period"). In other words, here, the user specifies at what time the captured (or registered) image is to be searched. Specifically, Tmin, which is the lower limit of the search target period, and Tmax, which is the upper limit of the search target period, are input. If there is no designation by the user, the entire period is set as the search target period.

Next, in step S603, a search key image is specified. Here, the specified search key image may be an image newly input from the scanner 32 or the like as described above, or may be an image selected from images stored in the database 301 in advance.

When the image is selected from among the images already stored, a display screen for specifying a search key image as shown in FIG. As shown in the figure, here, a list of search key images based on the images stored in the database 301 is displayed. At the same time, a character string that prompts the user to click an image to search for is displayed. When an appropriate image is clicked by the user, that image is designated as a search key image.

When the search target period and the search key image are specified, the initial setting of the search reference date T is performed in step S605. Here, the time information of the designated search key image, that is, the shooting date information or the registration date information Tkey given to the search key image is used as the search reference date Tkey.
(T ← Tkey).

In step S607, the search reference date T is set to Tmi which is the lower limit of the designated search target period.
It is determined whether it is smaller than n (T <Tmin?). By comparing the search reference date T with the lower limit Tmin of the search target period, it is determined whether or not the image search should be performed on an image past the search reference date T.

If the search reference date T is greater than Tmin, which is the lower limit of the search target period (“N” in step S607), in step S609, the image of the period preceding the search reference date T by a predetermined period Td is past. Is searched for. That is, the search is performed on the image having the time information of the period from the search reference date T to the date (T-Td) which is a predetermined period earlier.

Specifically, a specified search key image,
The difference from the image in the database 301 having the time information within the above-described period is obtained, and the similarity between the two is determined.
Then, for an image having a similarity greater than or equal to a predetermined value,
Once extracted as a "similar image". An image whose similarity is less than the predetermined value is not a “similar image” and is excluded from extraction.

When the search for the image having the time information from T to (T-Td) is completed, the search key image is reset in step S611. For example, among the “similar images” extracted in step S609, the image having the oldest time information is
A new search key image is set. The number of images is not limited to one, and a plurality of images may be set as new search key images in order from the oldest image.

Subsequently, in step S613, the search reference date T is updated. That is, in order to shift the period in which the search is performed (hereinafter referred to as “search period”) in the past, the search reference date T is shifted in the past by the predetermined period Td (T ← T−Td). Thereafter, step S60 is performed again.
It returns to the process of 7. Thereafter, the same processing using the new search key image is repeated for the shifted search period (steps S609 to S613).

If the search reference date T exceeds the lower limit Tmin of the search target period in step S607, that is, if T <Tmin, step S615 is performed.
Move to the processing of.

In step S615, the search reference date T is initialized again. That is, the time information Tkey of the search key image first specified by the user is substituted for the search reference date T (T ← Tkey). Then, in the subsequent processing, a search for an image having time information in the future from the search reference date T is performed.

In step S617, the search reference date T is
It is determined whether it is larger than Tmax which is the upper limit of the search target period (T> Tmax?). Here, by comparing the search reference date T with the upper limit Tmax of the search target period, it is determined whether or not the image search should be performed on images future from the reference date T.

The search reference date T is the upper limit Tmax of the search target period.
If it is smaller than (“N” in step S617), in step S619, a search is performed on the search reference date T and images in a period that is later than the search reference date T by a predetermined period Td in the future. That is, the search is performed on the image having the time information of the period from the search reference date T to the day (T + Td) advanced by the predetermined period.

More specifically, similarity determination is performed by calculating the difference between the search key image and the image having the time information of the search period, similarly to the processing performed in step S609. If the similarity of the target image is equal to or greater than a predetermined value, the image is extracted as a “similar image”. On the other hand, if it is less than the predetermined value, it is determined that the image is not a “similar image”.

Subsequently, in step S621, a search key image is reset. Specifically, among the “similar images” extracted in step S619, an image having the latest time information is newly set as a search key image. Note that a plurality of new key images may be set in order from the newest one instead of one.

When a new search key image is set, the search reference date T is updated in step S623. Here, in order to shift the search period in the future, the search reference date T is shifted in the future by a predetermined period Td (T ← T + Td). When the search reference date is updated in this way, the process returns to step S617. Thereafter, the same processing (steps S619 to S623) using the new search key image is repeated for the shifted search period.

In step S617, if the search reference date T exceeds the upper limit Tmax of the search target period, that is, if T> Tmax, the image search for the future is terminated. The process proceeds to S625.

In step S625, a search result is output within the search target period. That is, from the images extracted by the search processing in steps S609 and S619, an image having information within a designated period designated by the user is output as a search result. The output of the result may be displayed on the display device 12 or the printer 37.
May be printed out.

In this way, a new search key image is set as appropriate only by the user designating a desired search period and a search key image, and an image search is automatically performed based on the new search key image. Therefore, the operation performed by the user to obtain a desired image without omission is simple,
Convenience is improved.

Next, the case of searching for an image will be described in detail with reference to FIGS. 8 and 9 in accordance with the flowchart of the image data search processing of FIG.

FIG. 8 is a view for explaining a first specific example in the case of searching for an image according to the flowchart shown in FIG. Here, an image having time information between the lower limit Tmin of the search target period and the upper limit Tmax of the search target period is designated as the search key image. Therefore, first, a search is performed for an image having past time information from the search reference date T in the direction of arrow A, that is, from the search reference date (steps S607 to S613 in FIG. 6). Thereafter, a search is performed for images having time information in the direction of arrow B from the reference date T, that is, from the search reference date to the future (steps S617 to S623 in FIG. 6). This will be specifically described as follows.

First, using the specified search key image,
An image search is performed in a period (T to t4) that is a predetermined period Td before the search reference date T in the past. For example, when the predetermined period Td is one year, a search is performed for images captured in the past year from the shooting date T of the search key image. Then, an image similar to the search key image is extracted as a search result.

In order to search for an image that goes back in the past, the oldest image among the images extracted as a search result is set as a new search key image. Then, the search reference date T is also updated to t4. For this reason, based on the newly set search key image, an image search is performed next in a period (t4 to t3) that goes back from t4 to a predetermined period Td. Assuming that the predetermined period Td is one year, t4
From the last one year. Then, the oldest image among the searched images is set again as a new search key image.

In this manner, the setting of a new search key image and the shifting of the search period to the past are repeated in the same manner until the search target period lower limit Tmin is exceeded.
Here, the image search process in the period from t2 to t1 ends, and when the search reference date T is newly set to t1, the search reference date T exceeds the search target period lower limit Tmin, and goes to the past. The search processing ends.

Therefore, next, a search process is performed in the direction of arrow B from the time information T of the specified search key image. That is, a search process is performed on an image having time information that is later than the time information T of the search image (steps S617 to S623 in FIG. 6).

First, a predetermined period Td from T serving as a search reference date
Only during the period (T to t5) that goes to the future, the image search processing is performed. When the predetermined period T is one year, a search is performed for images that have been shot up to one year ahead of the shooting date T of the search key image. Then, an image similar to the search key image is extracted as a search result.

In order to search for a future image,
From among the images extracted as the search results, the image having the latest time information is set as a new search key image. Then, using the search key image, an image is searched for a predetermined period Td from t5 in a future period (from t5 to t6).

In this manner, the setting of a new search key image and the shift of the search period to the future are repeated in the same manner until the search target period upper limit Tmax is exceeded. Here, the image search process in the period from t5 to t6 is completed, and when the search reference date T is newly set to t6, the search reference date T exceeds the search target period upper limit Tmax, and it goes to the future. The search processing ends.

Finally, the search target period (Tmi
An image having time information within (n to Tmax) is output as a search result.

As described above, in the image search method according to the present embodiment, the search period is gradually shifted by a predetermined period toward the past and the future around the search reference date T. At the same time, the search key image is gradually updated to an old image or a new image. Therefore, a desired image up to the lower limit or the upper limit of the search target period can be searched without omission.

FIG. 9 is a diagram for explaining a second specific example in the case of searching for an image according to the flowchart shown in FIG. Here, an image having time information larger than the upper limit Tmax of the search target period is specified as the search key image. Therefore, unlike the example of FIG.
From to the search period. (Arrow A), that is, a search is performed for an image having time information in the past from the search reference date (steps S607 to S613 in FIG. 6), but from the search reference date T in the direction of arrow B, ie, from the search reference date to the future. Is not searched for the image having the time information (steps S617 to S623 in FIG. 6).

First, using the designated search key image,
An image search is performed in a period (T to t4) that is a predetermined period Td before the search reference date T in the past. Then, an image similar to the search key image is extracted as a search result.

Further, in order to search for an image that goes back in the past, the oldest image among the images extracted as a search result is set as a new search key image. Then, the search reference date T is also updated to t4. For this reason, based on the newly set search key image, an image search is performed next in a period (t4 to t3) that goes back from t4 to a predetermined period Td. Then, the oldest image among the searched images is set again as a new search key image.

In this manner, the setting of a new search key image and the shift of the search period to the past are repeated in the same manner until the search target period lower limit Tmin is exceeded.
Here, the image search process in the period from t2 to t1 ends, and when the search reference date T is newly set to t1, the search reference date T exceeds the search target period lower limit Tmin, and goes to the past. The search processing ends.

In this case, since the search reference date T is larger than the search target period upper limit Tmax, the search processing in the direction of arrow B is not performed. Therefore, the extraction of the image ends here, and the final search result is selected. Eventually,
Only images having time information within the search target period (Tmin to Tmax) are output as search results.

As described above, when the search reference date T is outside the search target period, even if the search reference date T is a period not including the search target period (in the figure, from T to t4 and from t4 to t3), the search target date T The search is performed while shifting by the predetermined period Td toward the period. This is the difference between Tmin and Tm
This is because it is necessary to sequentially set new search key images required for performing appropriate search processing in the search target period up to ax. Therefore, in the search target period, an appropriate image is extracted based on the search key image considering the passage of time.

FIG. 10 is a diagram for explaining the concept of updating the search key image. As shown in the figure, the search key image 1 is, for example, an image specified by clicking a third image from the left in FIG.
Therefore, first, an image search within a certain period based on the search key image 1 is performed.

Then, a plurality of similar images as shown in the next hierarchy are extracted as the search results. When searching for a past image, the oldest image is selected from the extracted similar images. Conversely, if a search is made for a future image, the most recent image is selected. Then, the selected image is set as a new search key image (search key image 2).

Then, an image search is performed based on search key image 2 which is a new search key image that has been set. At this time, the search target period is a period shifted in the past or the future.

Similarly, a new search key image is set as needed, and an image search process is performed based on the key image. All the images (search results) extracted in this manner are output as images similar to the search key image 1.

Therefore, according to such an image search method, a desired image can be extracted without omission. That is, even if the search period is long,
Even if the time difference between the search period and the specified search key image is large, for example, for a person, the change in age,
It can sufficiently cope with changes in body shape, hairstyle, and the like, and an accurate image search is performed. This eliminates the need for the user to select a plurality of representative search key images by predicting the search result in advance. In addition, the complexity of selecting a new search key image in consideration of the tendency of the image extracted by the original search key image can be avoided.

In FIG. 6, an image having time information in the past from the search reference date is searched first, and then an image having time information in the future is searched. However, the order is not limited. For example, the reverse may be performed, such as performing an image search for a past image after an image search for a future image.

Further, in steps S611 and S621 in FIG. 6, the image set as a new search key image is not limited to the image described in FIG.
For example, a plurality of new and old ones may be selected. In this case, a search process is performed for each selected image.

In the embodiment shown this time, a case where a person is designated as a search key image has been described as an example, but the search target is not limited to a person. Therefore, for example, the present invention can be applied to a case where an object such as a building is designated as a search key image and an image of the object is searched.

As described with reference to FIGS. 7, 8 and 9, etc., the search periods for performing image search are shifted by a predetermined period Td without overlapping. However, the present invention is not limited to such a case. For example, when the time information of an image set as a new search image is set as a new search reference date T, the search period is shifted while slightly overlapping. You may do so. Further, even if a new search key image is set without shifting from the beginning, the search may be repeatedly performed over the entirety each time.

The embodiments disclosed this time are to be considered in all respects as illustrative and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

[Brief description of the drawings]

FIG. 1 is an image retrieval apparatus 1 according to an embodiment of the present invention.
FIG. 10 is a diagram illustrating an appearance of a computer as an example of 00.

FIG. 2 is a functional block diagram showing a configuration of the computer shown in FIG.

FIG. 3 is a diagram for explaining a case where an image is input to and output from the image search device 100;

FIG. 4 is a flowchart showing a flow of processing in a control unit 300 of the image search device 100.

FIG. 5 is an image data registration process of FIG. 4 (step S40);
It is the flowchart which showed the detail of 5).

FIG. 6 is an image data search process of FIG. 4 (step S40);
It is the flowchart which showed the detail of 7).

FIG. 7 is a diagram illustrating an example in which a search key image designation screen is displayed on the display device 12.

FIG. 8 is a diagram for explaining a first specific example when searching for an image according to the flowchart shown in FIG. 6;

FIG. 9 is a diagram illustrating a second specific example in a case where an image is searched according to the flowchart illustrated in FIG. 6;

FIG. 10 is a diagram for explaining the concept of updating a search key image.

[Explanation of symbols]

11 computer main body, 12 display device, 13 magnetic tape device, 14 magnetic tape, 17 CD-ROM device, 18 CD-ROM, 20 CPU, 21 RO
M, 22 RAM, 23 hard disk drive, 31
Digital camera, 32 scanner, 33, 35 server, 37 printer, 300 control unit, 301 database.

 ────────────────────────────────────────────────── ─── Continuation of the front page (72) Kyoko Nakamura Inventor F-term in Osaka International Building Minolta Co., Ltd. 2-3-13 Azuchicho, Chuo-ku, Osaka 5B050 GA08 5B075 ND06 NK10 PP02 PP03 PP30 PQ02 PQ03 PR06 QM08

Claims (8)

[Claims] A search key image obtaining step of obtaining a search key image specified by a user; a search step of searching for a corresponding image from a database based on the obtained search key image; A setting step of setting a new search key image using a search result obtained by the step; and a re-search step of searching again for the corresponding image from the database based on the set new search key image. , A computer-readable recording medium on which an image search program for causing a computer to execute is stored. 2. The search step performs a search on a first group in the database, and the re-search step performs a search on a second group different from the first group in the database. A computer-readable recording medium on which the image search program according to claim 1 is searched for. 3. An image stored in the database is provided with time information, and the first group is a group of images having time information within a first period. The computer-readable recording medium recording the image search program according to claim 2, wherein the second group is a group of images having time information within the second period. 4. The image according to claim 3, wherein in the setting step, an image having time information close to the second period is set as a new search key image from among the images searched in the search step. A computer-readable recording medium recording a search program. 5. A search target period obtaining step for obtaining a search target period designated by a user, and among the images searched by the search step and the re-search step, the search target period within the obtained search target period is obtained. A computer-readable recording medium storing the image search program according to claim 3, wherein the computer executes a determining step of determining an image having time information as a search result. 6. A computer further executes a resetting step of setting a new search key image using a search result of the re-searching step, wherein the resetting step and the re-searching step are repeated a predetermined number of times A computer-readable recording medium storing the image search program according to claim 1. 7. A search key image obtaining step of obtaining a search key image specified by a user; a search step of searching a database for a corresponding image based on the obtained search key image; A setting step of setting a new search key image using a search result obtained by the step; and a re-search step of searching again for the corresponding image from the database based on the set new search key image. An image search method, including: 8. A search key image obtaining means for obtaining a search key image specified by a user; a search means for searching a database for a corresponding image based on the obtained search key image; Setting means for setting a new search key image using a search result by the means, and re-search means for searching again the corresponding image from the database based on the set new search key image. An image search device, including: